Scan to contactWill this consume tokens?
No. The check only asks the endpoint for its model list and does not run a chat completion.
shield_lockKey Safety Check
Check Key Availability
Check whether your relay endpoint and key are usable, OpenAI-compatible, and which models it exposes. The key is only used for this check.
lockNo key storage
We do not store your API Key. This check only calls /models once and does not consume tokens.
tokenNo token burn
No token-consuming prompt is sent during the check.
view_listModel exposure
See which model names the endpoint exposes.
Check Configuration
Paste endpoint and a test key
history
key
shield_lockWe do not store your API Key. This check only calls /models once and does not consume tokens.
Live Result
Check Result
radarWaiting for a key check
Results appear here after the /models handshake
What this key checker verifies
This tool focuses on the endpoint and key handshake before you spend tokens on generation requests.
vpn_key
Authentication
Checks whether the key can access the endpoint without returning 401 or permission errors.
hub
OpenAI compatibility
Calls the model list endpoint to confirm whether the relay follows an OpenAI-compatible shape.
view_list
Model exposure
Lists available model names so you can avoid model_not_found and alias mismatch issues.
lock
Privacy boundary
The check only calls /models and does not send a prompt or store the key in the page result.
FAQ
helpQuick answersKey checker FAQ
Why can authentication pass but a model still fail?
Some relays accept the key but expose only part of their model list, or use custom model aliases.
What should I do after a failed check?
Verify the endpoint and key belong together, inspect the returned error, then compare another relay if the issue persists.